mirror of
https://github.com/jambonz/speech-utils.git
synced 2026-07-04 19:31:49 +00:00
wip
This commit is contained in:
Hoan HL
+220
@@ -256,6 +256,226 @@ test('Google Gemini TTS synth tests', async(t) => {
|
||||
client.quit();
|
||||
});
|
||||
|
||||
test('Google TTS streaming tests (!JAMBONES_DISABLE_TTS_STREAMING)', async(t) => {
|
||||
// Ensure streaming is enabled (default behavior)
|
||||
delete process.env.JAMBONES_DISABLE_TTS_STREAMING;
|
||||
|
||||
// Clear require cache to reload config with new env var
|
||||
delete require.cache[require.resolve('../lib/config')];
|
||||
delete require.cache[require.resolve('../lib/synth-audio')];
|
||||
delete require.cache[require.resolve('..')];
|
||||
|
||||
const fn = require('..');
|
||||
const {synthAudio, client} = fn(opts, logger);
|
||||
|
||||
if (!process.env.GCP_FILE && !process.env.GCP_JSON_KEY) {
|
||||
t.pass('skipping Google TTS streaming tests since neither GCP_FILE nor GCP_JSON_KEY provided');
|
||||
return t.end();
|
||||
}
|
||||
|
||||
try {
|
||||
const str = process.env.GCP_JSON_KEY || fs.readFileSync(process.env.GCP_FILE);
|
||||
const creds = JSON.parse(str);
|
||||
const geminiModel = process.env.GCP_GEMINI_TTS_MODEL || 'gemini-2.5-flash-tts';
|
||||
|
||||
// Test 1: Standard voice streaming (use_live_api=0)
|
||||
let result = await synthAudio(stats, {
|
||||
vendor: 'google',
|
||||
credentials: {
|
||||
credentials: {
|
||||
client_email: creds.client_email,
|
||||
private_key: creds.private_key,
|
||||
},
|
||||
},
|
||||
language: 'en-US',
|
||||
voice: 'en-US-Wavenet-D',
|
||||
gender: 'MALE',
|
||||
text: 'This is a test of standard voice streaming.',
|
||||
disableTtsCache: true
|
||||
});
|
||||
t.ok(result.filePath.startsWith('say:'), 'Standard voice returns streaming say: path');
|
||||
t.ok(result.filePath.includes('vendor=google'), 'Standard voice streaming path contains vendor=google');
|
||||
t.ok(result.filePath.includes('use_live_api=0'), 'Standard voice uses use_live_api=0');
|
||||
t.ok(result.filePath.includes('voice=en-US-Wavenet-D'), 'Standard voice streaming path contains voice');
|
||||
|
||||
// Test 2: HD voice streaming (use_live_api=1)
|
||||
result = await synthAudio(stats, {
|
||||
vendor: 'google',
|
||||
credentials: {
|
||||
credentials: {
|
||||
client_email: creds.client_email,
|
||||
private_key: creds.private_key,
|
||||
},
|
||||
},
|
||||
language: 'en-US',
|
||||
voice: 'en-US-Chirp3-HD-Charon',
|
||||
text: 'This is a test of HD voice streaming.',
|
||||
disableTtsCache: true
|
||||
});
|
||||
t.ok(result.filePath.startsWith('say:'), 'HD voice returns streaming say: path');
|
||||
t.ok(result.filePath.includes('vendor=google'), 'HD voice streaming path contains vendor=google');
|
||||
t.ok(result.filePath.includes('use_live_api=1'), 'HD voice uses use_live_api=1 (Live API)');
|
||||
t.ok(result.filePath.includes('voice=en-US-Chirp3-HD-Charon'), 'HD voice streaming path contains voice');
|
||||
|
||||
// Test 3: Gemini TTS streaming (use_live_api=1)
|
||||
result = await synthAudio(stats, {
|
||||
vendor: 'google',
|
||||
credentials: {
|
||||
credentials: {
|
||||
client_email: creds.client_email,
|
||||
private_key: creds.private_key,
|
||||
},
|
||||
},
|
||||
language: 'en-US',
|
||||
voice: 'Kore',
|
||||
model: geminiModel,
|
||||
text: 'This is a test of Gemini TTS streaming.',
|
||||
instructions: 'Speak naturally.',
|
||||
disableTtsCache: true
|
||||
});
|
||||
t.ok(result.filePath.startsWith('say:'), 'Gemini TTS returns streaming say: path');
|
||||
t.ok(result.filePath.includes('vendor=google'), 'Gemini TTS streaming path contains vendor=google');
|
||||
t.ok(result.filePath.includes('use_live_api=1'), 'Gemini TTS uses use_live_api=1 (Live API)');
|
||||
t.ok(result.filePath.includes(`model_name=${geminiModel}`), 'Gemini TTS streaming path contains model_name');
|
||||
t.ok(result.filePath.includes('prompt=Speak naturally.'), 'Gemini TTS streaming path contains prompt');
|
||||
|
||||
// Test 4: Gemini TTS with SSML stripping in streaming mode
|
||||
result = await synthAudio(stats, {
|
||||
vendor: 'google',
|
||||
credentials: {
|
||||
credentials: {
|
||||
client_email: creds.client_email,
|
||||
private_key: creds.private_key,
|
||||
},
|
||||
},
|
||||
language: 'en-US',
|
||||
voice: 'Leda',
|
||||
model: geminiModel,
|
||||
text: '<speak>This SSML should be stripped.</speak>',
|
||||
instructions: 'Speak naturally.',
|
||||
disableTtsCache: true
|
||||
});
|
||||
t.ok(result.filePath.startsWith('say:'), 'Gemini TTS with SSML returns streaming say: path');
|
||||
t.ok(!result.filePath.includes('<speak>'), 'SSML tags are stripped from streaming path');
|
||||
t.ok(result.filePath.includes('This SSML should be stripped.'), 'Text content is preserved after SSML stripping');
|
||||
|
||||
// Test 5: Gemini TTS with prompt containing special characters
|
||||
result = await synthAudio(stats, {
|
||||
vendor: 'google',
|
||||
credentials: {
|
||||
credentials: {
|
||||
client_email: creds.client_email,
|
||||
private_key: creds.private_key,
|
||||
},
|
||||
},
|
||||
language: 'en-US',
|
||||
voice: 'Kore',
|
||||
model: geminiModel,
|
||||
text: 'Testing special characters in prompt.',
|
||||
options: { prompt: 'Speak in a warm, friendly tone' },
|
||||
disableTtsCache: true
|
||||
});
|
||||
t.ok(result.filePath.startsWith('say:'), 'Gemini TTS with special chars returns streaming say: path');
|
||||
// Commas in prompt should be replaced with semicolons
|
||||
t.ok(result.filePath.includes('prompt=Speak in a warm; friendly tone'), 'Commas in prompt are escaped to semicolons');
|
||||
|
||||
} catch (err) {
|
||||
console.error(err);
|
||||
t.end(err);
|
||||
}
|
||||
client.quit();
|
||||
});
|
||||
|
||||
test('Google TTS non-streaming tests (JAMBONES_DISABLE_TTS_STREAMING=true)', async(t) => {
|
||||
// Enable streaming disable flag
|
||||
process.env.JAMBONES_DISABLE_TTS_STREAMING = 'true';
|
||||
|
||||
// Clear require cache to reload config with new env var
|
||||
delete require.cache[require.resolve('../lib/config')];
|
||||
delete require.cache[require.resolve('../lib/synth-audio')];
|
||||
delete require.cache[require.resolve('..')];
|
||||
|
||||
const fn = require('..');
|
||||
const {synthAudio, client} = fn(opts, logger);
|
||||
|
||||
if (!process.env.GCP_FILE && !process.env.GCP_JSON_KEY) {
|
||||
t.pass('skipping Google TTS non-streaming tests since neither GCP_FILE nor GCP_JSON_KEY provided');
|
||||
delete process.env.JAMBONES_DISABLE_TTS_STREAMING;
|
||||
return t.end();
|
||||
}
|
||||
|
||||
try {
|
||||
const str = process.env.GCP_JSON_KEY || fs.readFileSync(process.env.GCP_FILE);
|
||||
const creds = JSON.parse(str);
|
||||
const geminiModel = process.env.GCP_GEMINI_TTS_MODEL || 'gemini-2.5-flash-tts';
|
||||
|
||||
// Test 1: Standard voice falls back to non-streaming API
|
||||
let result = await synthAudio(stats, {
|
||||
vendor: 'google',
|
||||
credentials: {
|
||||
credentials: {
|
||||
client_email: creds.client_email,
|
||||
private_key: creds.private_key,
|
||||
},
|
||||
},
|
||||
language: 'en-US',
|
||||
voice: 'en-US-Wavenet-D',
|
||||
gender: 'MALE',
|
||||
text: 'This is a test with streaming disabled.',
|
||||
disableTtsCache: true
|
||||
});
|
||||
t.ok(!result.filePath.startsWith('say:'), 'Standard voice does NOT return streaming say: path when disabled');
|
||||
t.ok(result.filePath.endsWith('.mp3'), 'Standard voice returns mp3 file path');
|
||||
|
||||
// Test 2: HD voice falls back to non-streaming API
|
||||
result = await synthAudio(stats, {
|
||||
vendor: 'google',
|
||||
credentials: {
|
||||
credentials: {
|
||||
client_email: creds.client_email,
|
||||
private_key: creds.private_key,
|
||||
},
|
||||
},
|
||||
language: 'en-US',
|
||||
voice: 'en-US-Chirp3-HD-Charon',
|
||||
text: 'This is a test of HD voice with streaming disabled.',
|
||||
disableTtsCache: true
|
||||
});
|
||||
t.ok(!result.filePath.startsWith('say:'), 'HD voice does NOT return streaming say: path when disabled');
|
||||
t.ok(result.filePath.endsWith('.mp3'), 'HD voice returns mp3 file path');
|
||||
|
||||
// Test 3: Gemini TTS falls back to non-streaming API
|
||||
result = await synthAudio(stats, {
|
||||
vendor: 'google',
|
||||
credentials: {
|
||||
credentials: {
|
||||
client_email: creds.client_email,
|
||||
private_key: creds.private_key,
|
||||
},
|
||||
},
|
||||
language: 'en-US',
|
||||
voice: 'Kore',
|
||||
model: geminiModel,
|
||||
text: 'This is a test of Gemini TTS with streaming disabled.',
|
||||
instructions: 'Speak naturally.',
|
||||
disableTtsCache: true
|
||||
});
|
||||
t.ok(!result.filePath.startsWith('say:'), 'Gemini TTS does NOT return streaming say: path when disabled');
|
||||
t.ok(result.filePath.endsWith('.mp3'), 'Gemini TTS returns mp3 file path');
|
||||
|
||||
} catch (err) {
|
||||
console.error(err);
|
||||
t.end(err);
|
||||
} finally {
|
||||
// Clean up: restore default behavior
|
||||
delete process.env.JAMBONES_DISABLE_TTS_STREAMING;
|
||||
delete require.cache[require.resolve('../lib/config')];
|
||||
delete require.cache[require.resolve('../lib/synth-audio')];
|
||||
delete require.cache[require.resolve('..')];
|
||||
}
|
||||
client.quit();
|
||||
});
|
||||
|
||||
test('AWS speech synth tests', async(t) => {
|
||||
const fn = require('..');
|
||||
const {synthAudio, client} = fn(opts, logger);
|
||||
|
||||
Reference in New Issue
Block a user